BACKGROUND: Pulsinelli et al developed a kind of rat models with four-vessel occlusion-induced global cerebral ischemia. Because the histo-pathological changes and severe cerebral ischemia reproducibility of this model are good and the stability of this model in circulation respiration is superior to that of other models, so four-vessel occlusion method has become a classic modeling method for global cerebral ischemia model. This model has been improved in some laboratories to meet different requirements in different studies. OBJECTIVE: To establish a highly reproducible rat model of reversible forebrain ischemia by modifying four-vessel occlusion model introduced by Pulsinelli et al, and to investigate its neurophysiological and pathological changes and the characteristics of modified operation. DESIGN: Completely randomized grouping, controlled trial. SETTING: Department of Anesthesiology,Changhai Hospital, Second Military Medical University of Chinese PLA. MATERIALS: A total of 65 male healthy SD rats, weighing 250-300 g, were provided by the Experimental Animal Center of Second Military Medical University of Chinese PLA. VSM hemodynamic monitor and temperature monitor (Thermal ert TH-5, U.S.A) were used.METHODS: The trial was conducted in the Department of Anesthesiology, Changhai Hospital, Second Military Medical University of Chinese PLA from January 2005 to March 2006. ① Experimental grouping: Sixty-five rats were randomly divided into the following 7 groups: sham-operation group (n =9): given the same operation, without occlusion of vessels; 5 minutes ischemia group (n =9): ischemia 5 minutes and reperfusion 72 hours; 10 minutes ischemia group (n =8): ischemia 10 minutes and reperfusion 72 hours; 15 minutes ischemia group (n =9): ischemia 15 minutes and reperfusion 72 hours; 20 minutes ischemia group (n =8): ischemia 20 minutes and reperfusion 72 hours; 30 minutes ischemia group (n =7); ischemia 30 minutes and reperfusion 72 hours; ischemia control group (n =15): ischemia 15 minutes and reperfusion 180 minutes. ② Preparation of the model of global cerebral ischemia: Four-vessel occlusion- induced global cerebral ischemia rat models were modified, I.e. Bilateral vertebral arteries could be electrocauterized and blocked, and bilateral common carotid arteries were enclosed with 10-0 suture loosely. On the second day, keeping the animal awake, the suture was tightened and kept tense to block the blood flow of bilateral common carotid artery. After certain duration of ischemia, the suture was cut off and drawn out, thus the reperfusion of bilateral common carotid artery was resumed. ③Observation of physiological indexes: VSM hemodynamic monitor and temperature monitor (Thermal ert TH-5,U.S.A) were used to record and measure the changes of blood pressure, rectal temperature, brain temperature ,arterial blood gas and other physiological indexes of the rats in the control group before, 5 and 15 minutes after ischemia and 10,30,60,120 and 180 minutes after reperfusion. ④Preparation of brain tissue pathological samples: Except for ischemia control group, rats of other groups were anesthetized and their brain tissues were harvested and subjected to haematoxylin and eosin staining at 72 hours after reperfusion. ⑤ rnEvaluation of pathological change of brain tissue: The coronal plane of anterior commissure of cerebrum was used to evaluate corpora striatum, and the coronal plane of anterior hippocampus was used to evaluate hippocampal CA1/2 region, CA3 region and CA4 region, subiculum, superior pyramidal lobe and inferior pyramidal lobe of dentate gyrus as well as neocortex. Irreversible neuronal damage included pyknotic cells with eosinophilic cytoplasm and trachychromatic nucleus, homogenous cytoplasm and naked nucleus. Neurons without the above changes were considered to be normal. The number of normal neurons in the above-mentioned brain regions was counted under the microscope. MAIN OUTCOME MEASURES: ① The changes of blood pressure, rectal temperature, brain temperature, arterial blood gas and other physiological indexes of rats before, 5 and 15 minutes after ischemia and 10,30,60,120 and 182 minutes after reperfusion. ②The number of normal neurons in hippocampal CA1/2 region, CA3 region, CA4 region, subiculum, superior pyramidal lobe and inferior pyramidal lobe of dentate gyrus as well as neocortex at 72 hours of ischemia. RESULTS: Sixty-five SD rats were involved in the result analysis. ①Changes in the physiological indexes of rats during cerebral ischemia/reperfusion: Changes in arterial blood pressure: arterial blood pressure at 5 minutes after ischemia was significantly higher than that before ischemia [(132.82±12.38) mm Hg vs. (118.58±5.85) mm Hg, 1 mm Hg=0.133 kPa, P ＜ 0.01], and there was no significant difference in arterial blood pressure between at ischemia 15 minutes, reperfusion 10,30,60,120 and 180 minutes and before ischemia (P ＞ 0.05); Changes in arterial blood gas: pH value at 15 minutes after ischemia was significantly higher than that before ischemia (7.56±0.05 vs. 7.47±0.09, P ＜ 0.01 ), PaO2 value at 15 minutes after ischemia was significantly higher than that beischemia [(1.71±0.04) kPa vs. (1.56±0.08) kPa, P＜ 0.01], and PaCO2 value at 5 and 15 minutes after ischemia was significantly lower than that before ischemia [(0.44±0.04), (0.37±0.04), (0.57± 0.04) kPa, P ＜ 0.01];Changes in rectal temperature and brain temperature: Brain temperature at 5 and 10 minutes after ischemia and at 10 and 30 minutes after reperfusion was significantly lower than that before ischemia [(37.01±0.43), (36.44±0.35), (36.81 ±0.45), (37.23±0.24), (37.74±0.28) ℃,P＜ 0.01], and rectal temperature at 10 minutes after reperfusion was significantly lower than that before ischemia [(36.76±0.21), (37.37± 0.19) ℃, P ＜ 0.01]. ② Changes in the normal neurons in different encephalic regions of rats at different time after ischemia and 72 hours after reperfusion: At 5,10,15,20 and 30 minutes after ischemia, the number of normal neurons in subiculum was 86%, 73%, 11%, 8% ,6% of that in the Sham-operation, respectively (P ＜ 0.05-0.01), the number of normal neurons in intal segment of hippocampal CA1/2 region was 74% ,53% ,8%, 5%, 5% of that in the sham-operation group (P ＜ 0.01 ),the number of normal neurons in middle segment of hippocampal CA1/2 was 73%,48%, 11%,8%,6% of that in the sham-operation group, respectively (P ＜ 0.01), and the number of normal neurons in extal segment of hippocampal CA1/2 region was 90% ,72% ,37%, 31% ,22% of that in the sham-operation group, respectively (P ＜ 0.01 ); At 10,15,20 and 30 minutes after ischemia, the number of normal neurons in hippocampal CA3 region was 80% ,67% ,47% ,40% of that in the sham-operation group respectively (P ＜ 0.01 ), the number of normal neurons in hippocampal CA4 region was 79%, 70%, 60% ,43% of that in the sham-operation group, respectively (P ＜ 0.01 ) and the number of normal neurons in superior pyramidal lobe of dentate gyrus was 85%,75%, 59%, 52% of that in the sham-operation group, respectively (P ＜ 0.05,0.01). At 15,20 and 30 minutes after ischemia, the number of normal neurons in inferior pyramidal lobe of dentate gyrus was 80% ,70% ,59% of that in the sham-operation group (P ＜ 0.01 ) and the number of normal neurons in the neocortex was 79%, 72%,61% of that in the sham-operation group (P ＜ 0.05,0.01).CONCLUSION : To construct cerebral ischemia rat model with modified four-vessel occlusion can cause severe cerebral ischemia.lt has a few influences on arterial blood pressure and arterial blood gas of model rat and is easy for blocking and opening carotid artery system.